Decision Support for Declarative Artifact-Centric Process Models

Guiding Knowledge Workers Under Dynamic Contexts

“…Two works are especially relevant for our work [27,29]. In [27], MDPs are used to optimize the execution of Product Data Models (PDM).…”

“…Then, at any state of the process, based on the optimization objective such as minimizing completion time or costs, a next best step recommendation that, optimizes the objective globally, is provided. In [29], the authors translate a declarative artifact-centric process model into a MDP. Using the MDP, authors provide decision support on planning of the tasks in the process to support efficient information gathering throughout the process.…”

“…There are several approaches that propose support for KiPs [5,11,12,15,17,21,23,[27][28][29][30]. These approaches focus on guiding KiPs with some machine learning/artificial intelligence(AI) techniques based on historical information [11,15,21,30], modeling and planning of KiPs using AI (or automated planning) techniques and Markov Decision Process (MDP) [12,23,[27][28][29], and optimization of declarative processes [5,17]. KiPs are hardly repeatable [7], therefore useful historical information does not always exist.…”

“…KiPs are hardly repeatable [7], therefore useful historical information does not always exist. Different from existing approaches [5,11,12,15,17,21,[27][28][29][30], we focus on guiding knowledge workers on such cases where no historical data is present.…”

“…Also, we focus on information gathering aspect of tasks when prioritizing them, which is also not the focus of the existing approaches in [5,17]. Lastly, to the best of our knowledge, the notion of dynamic context and its integration for guiding knowledge workers in executing flexible KiPs, has not been investigated in the existing work [5,11,12,15,17,18,23,[27][28][29][30].…”